Optical Identification of the ASCA Large Sky Survey

71 Pages·0.47 MB·English

by  M. Akiyama

Optical Identification of the ASCA Large Sky Survey
Masayuki Akiyama1,2,10,11,12, Kouji Ohta2,10,11,Toru Yamada3,11, Nobunari Kashikawa4,
Masafumi Yagi4, Wataru Kawasaki5,12, Masaaki Sakano6,12, Takeshi Tsuru6, Yoshihiro
Ueda7,10, Tadayuki Takahashi7, Ingo Lehmann8,10, Gu¨nther Hasinger8, and Wolfgang
0
0 Voges9
0
2

n
a Received ; accepted
J

7
1

1 Revised Oct. 12th
v
9
8
2
1
0 1SUBARU Telescope, National Astronomical Observatory of Japan, 650 North A’ohoku
0
0
Place, Hilo, HI, 96720, U.S.A
/
h
p 2Department of Astronomy, Kyoto University, Kyoto 606-8502, Japan
-
o
r 3Astronomical Institute, Tohoku University, Sendai 980-8578, Japan
t
s
a 4National Astronomical Observatory of Japan, Mitaka, Tokyo 181-8588, Japan
:
v
i 5Department of Astronomy, University of Tokyo, Tokyo 113-8658, Japan
X
ar 6Department of Physics, Kyoto University, Kyoto 606-8502, Japan
7Institute of Space and Astronautical Science, Kanagawa 229-8510, Japan
8Astrophysikalisches Institut Potsdam, An der Sternwarte 16, 14482 Potsdam, Germany
9MPI fu¨r extraterrestrische Physik, Postfach 1603, 85740 Garching, Germany
10Visiting Astronomer, German-Spanish Astronomical Centre, Calar Alto, operated by
the Max-Plank-Institute for Astronomy, Heidelberg, jointly with the Spanish National
Commission for Astronomy
11Visiting Astronomer, University of Hawaii Observatory.
12Research Fellow of the Japan Society for the Promotion of Science.
– 2 –
ABSTRACT
We present results of optical identifications of the X-ray sources detected in the
ASCA Large Sky Survey. Optical spectroscopic observations were done for 34
X-ray sources which were detected with the SIS in the 2–7 keV band above 3.5
σ. The flux limit corresponds to ∼ 1 ×10−13 erg cm−2 s−1 in the 2–10 keV
band. The sources are identified with 30 AGNs, 2 clusters of galaxies, and 1
galactic star. Only 1 source is still unidentified.
All of the X-ray sources that have a hard X-ray spectrum with an apparent
photon index of smaller than 1 in the 0.7–10 keV band are identified with
narrow-line or weak-broad-line AGNs at redshifts smaller than 0.5. This fact
supports the idea that absorbed X-ray spectra of narrow-line and weak-broad-
line AGNs make the Cosmic X-ray Background (CXB) spectrum harder in the
hard X-ray band than that of a broad-line AGN, which is the main contributor
in the soft X-ray band. Assuming their intrinsic spectra are same as a broad-line
AGN (a power-law model with a photon index of 1.7), their X-ray spectra are
fitted with hydrogen column densities of logN (cm−2) = 22 ∼ 23 at the object’s
H
redshift. On the other hand, X-ray spectra of the other AGNs are consistent
with that of a nearby type 1 Seyfert. In the sample, four high-redshift luminous
broad-line AGNs show a hard X-ray spectrum with an apparent photon index
of 1.3±0.3. The hardness may be explained by the reflection component of a
type 1 Seyfert. The hard X-ray spectra may also be explained by absorption
with logN (cm−2) = 22 ∼ 23 at the object’s redshift, if we assume an intrinsic
H
photon index of 1.7. The origin of the hardness is not clear yet.
Based on the logN-logS relations of each population, contributions to the
CXB in the 2–10 keV band are estimated to be 9% for less-absorbed AGNs
(logN (cm−2) < 22) including the four high-redshift broad-line AGNs with a
H
– 3 –
hard X-ray spectrum, 4% for absorbed AGNs (22 < logN (cm−2) < 23, without
H
the four hard broad-line AGNs), and 1% for clusters of galaxies in the flux
range from 3×10−11 erg cm−2 s−1 to 2×10−13 erg cm−2 s−1 . If the four hard
broad-line AGNs are included in the absorbed AGNs, the contribution of the
absorbed AGNs to the CXB is estimated to be 6%.
In optical spectra, there is no high-redshift luminous cousin of a narrow-line
AGN in our sample. The redshift distribution of the absorbed AGNs are limited
below z = 0.5 excluding the four hard broad-line AGNs, in contrast to the
existence of 15 less-absorbed AGNs above z = 0.5. The redshift distribution
of the absorbed AGNs suggests a deficiency of AGNs with column densities
of logN (cm−2) = 22 to 23 in the redshift range between 0.5 and 2, or in the
H
X-ray luminosity range larger than 1044 erg s−1, or both. If the large column
densities of the four hard broad-line AGNs are real, they could complement the
deficiency of X-ray absorbed luminous high-redshift AGNs.
Subject headings: surveys — galaxies: active — quasars: general —
X-rays:galaxies — diffuse radiation
– 4 –
1. Introduction
Since the discovery of the cosmic X-ray background (CXB) by Giacconi et al. (1962) in the
2–6 keV band, many efforts have been made to understand the origin of the CXB. Recently
in ROSAT deep surveys, 70 – 80% of the CXB in the 0.5–2 keV band has been resolved into
discrete sources at a flux limit of 1×10−15 erg cm−2 s−1 (Hasinger et al. 1998). Within the
flux level of the deep survey, broad-line AGNs are the dominant population and the main
contributor to the CXB in the 0.5–2 keV band. On the other hand, in the harder 2–10 keV
∼
band, only 3% of the CXB was resolved into discrete sources (Piccinotti et al. 1982)
before ASCA surveys. Broad-line AGNs have X-ray power-law spectra with a photon index
of Γ = 1.7 in the 2–10 keV band (Turner & Pounds 1989) which are significantly softer than
that of the CXB in that band (Γ = 1.4 ∼ 1.5; Gendreau et al. 1995; Ishisaki et al. 1999),
thus there must be objects which have harder X-ray spectra than nearby broad-line AGNs
and contribute significantly to the CXB in the hard band.
Resolving the hard X-ray sky is a direct way to reveal the nature of X-ray sources in
the hard band. ASCA GIS observations of three ROSAT Deep PSPC Fields down to the
5×10−14 erg cm−2 s−1 have been done so far (Georgantopoulos et al. 1997; Boyle et al.
1998a). However, because of the large positional uncertainties of the GIS-selected X-ray
sources, a large fraction of hard X-ray selected sources is still unidentified; the nature of
hard X-ray sources and the difference from the ROSAT selected objects are still unclear.
Based on the unified scheme of AGNs, absorbed AGNs are proposed as candidates for
the hard X-ray sources (Comastri et al. 1995; Madau, Ghisellini, & Fabian 1994). Because
of absorption of soft X-ray photons by obscuring material, they have harder X-ray spectra
than type 1 AGNs. To reproduce the CXB spectrum, it is argued that there are around
three times more absorbed AGNs than non-absorbed AGNs in the universe and absorbed
AGNs dominate the CXB above 2 keV. In consequence of the assumption, the existence
– 5 –
of absorbed narrow-line QSO (so called type 2 QSO) is expected. From optical follow-up
observations of X-ray surveys in the soft and hard bands, several luminous narrow-line
AGNs have been found at intermediate to high redshift universe (e.g., Stocke et al. 1982;
Almaini et al. 1995; Ohta et al. 1996; Boyle et al. 1998b; Barcons et al. 1998; Schmidt et
al. 1998). On the other hand, existence of red broad-line QSOs whose red color suggests
absorption to its nucleus was reported from a radio survey (Webster et al. 1995). Such
a population was also identified in ROSAT surveys (Kim & Elvis 1999) and Beppo-SAX
surveys (Fiore et al. 1999), but X-ray spectra, number densities and contributions to the
CXB of both of these narrow-line and red broad-line QSOs are not clear.
To reveal the nature of X-ray sources in the hard band, studies on a well-defined
sample are important. We are now conducting an unbiased large and deep survey with
ASCA in the region near the north Galactic pole, i.e., ASCA Large Sky Survey (hereafter
LSS; Inoue et al. 1996; Ueda 1996; Ueda et al. 1998; Ueda et al. 1999a (Paper I)). The flux
limit of the LSS (∼ 1 ×10−13 erg cm−2 s−1 in the 2–10 keV band) is 100 times deeper
than the HEAO1 A2 survey, which was the deepest systematic survey in the hard band
(Piccinotti et al. 1982) before ASCA. We have surveyed 7.0 deg2 and 5.4 deg2 with the
GIS and the SIS detectors, respectively. Combining the data from the GIS and the SIS, we
detected 44 sources in the 2–10 keV band with the following criteria: 1) the significance
of summed count rate of the GIS and the SIS should exceed 4.5, and 2) the significance
of either the GIS or the SIS should also exceed 3.5 (Paper I). They correspond to 20 –
30% of the CXB in this band. The advantages of the sample are 1) the survey area as
a function of limiting count rates was determined well by simulations, 2) accurate X-ray
positions were determined by using the SIS whose resolution is higher than the GIS and by
correcting the temperature-dependent misalignment between the focal plane detectors and
the attitude sensors, and 3) the X-ray spectrum of each source was determined by fitting
power-law model with the Galactic absorption to the GIS and the SIS data simultaneously
– 6 –
in the 0.7–10 keV band (Paper I). Determining their X-ray spectra is important not only
to know their X-ray properties but also to know the flux limit of the survey, because the
limit varies with the X-ray spectrum of each source. The average of the apparent photon
index of the 36 X-ray sources detected in the flux range between 0.8 ×10−13 and 4 ×10−13
erg cm−2 s−1 is Γ = 1.49±0.1 (Paper I), which is significantly harder than the spectra of
X-ray sources detected in shallower surveys in the 2–10 keV band and close to that of the
CXB. Identification of these sources is clearly important to understand the nature of hard
X-ray sources and the origin of the CXB.
In this paper, we report results of optical identifications of X-ray sources detected
in the hard band with the SIS in the ASCA LSS. The sample definition and selections
of optical candidates are discussed in Section 2, results of optical spectroscopy for the
selected candidates and reliability of the identifications are presented in Section 3, and
optical and X-ray spectroscopic properties of identified objects are described in Section 4.
The contribution of the population to the CXB and the redshift distribution of the AGNs
are discussed in Section 5 and 6, respectively. In Section 7, we present multi-wavelength
properties of the identified AGNs. Throughout this paper, we use q = 0.5 and H = 50 km
0 0
s−1 Mpc−1. We call each X-ray source with the exact name and the identification number,
like AX J132032+3326(227), for convenience.
2. Observations
2.1. The X-ray Survey Observations and the Sample Definition
To minimize the effect of the galactic absorption and contamination from bright X-ray
sources, the survey area was defined as a continuous region of ∼ 5 degree2 near the north
Galactic pole, centered at α=13h14m, δ=31◦30′ (J2000). Seventy-six pointings of the
– 7 –
survey-observations were done from Dec. 1993 to Jul. 1995. These exposures were designed
to evenly cover the whole survey region with a 20 ksec effective exposure of the SIS. The
source detections were done with the SIS data in the 0.7–7 keV, 0.7–2 keV, and 2–7 keV
bands and the GIS data in the 0.7–7 keV, 0.7–2 keV, and 2–10 keV bands. For details of
the survey observation, source extraction, and spectral fitting, including the survey region
and the position of the sources on the sky, see Paper I.
In this paper, we concentrate on the 34 X-ray sources detected with the SIS in the
2–7 keV band above 3.5σ (hereafter, the SIS 2–7 keV 3.5σ sample). Table 1 shows the
survey area as a function of limiting count rates for the sample. The typical and the deepest
limiting count rates of the sample are 2 counts ksec−1 and 1.2 counts ksec−1, respectively.
They correspond to 1.8×10−13 erg cm−2 s−1 and 1.1×10−13 erg cm−2 s−1 in the 2–10 keV
band for an X-ray source with a power-law spectrum with a photon index of 1.7. All
X-ray sources in the SIS 2–7 keV 3.5σ sample have the significance level larger than 4.5 in
summed count rate of the GIS and the SIS in both of the 0.7–7 keV and 2–7 keV bands and
′
the positional uncertainties of such sources were estimated to be 0.6 in radius with the 90%
confidence level (Paper I). The number of spurious sources was estimated to be at most
a few percent (Paper I), thus less than 1 spurious source was expected to be in the SIS
2–7 keV 3.5σ sample.
EDITOR: PLACE TABLE 1 HERE.
2.2. Optical Imaging Observations and Selections of Optical Counterpart
Candidates
AGNs are the most plausible optical counterparts for the majority of the X-ray sources, thus
at least we have to reach the optical flux limit which is converted from the deepest X-ray
– 8 –
flux limit based on X-ray-to-optical flux ratio of AGN. If we assume power-law spectra
with an X-ray photon index of 1.7 and an optical energy index of −0.5 together with the
X-ray-to-optical flux ratio of AGNs identified in the ROSAT Deep Survey (Schmidt et al.
1998), the expected optical magnitude for the optical counterpart of an X-ray source with
1.1×10−13 erg cm−2 s−1 in the 2–10 keV band is R = 16 ∼ 21 mag.
Candidates of optical counterparts were mainly selected from the APM catalog which is
obtained from scans of glass copies of the Palomar Observatory Sky Survey plate (McMahon
∼
et al. 1992). However, the limiting magnitude of the catalog is R = 20 mag and is not deep
enough to pick up optical counterparts for faint X-ray sources. To complement the depth
of the data, we made imaging observations of the LSS region at the KISO 1.05m Schmidt
telescope in March and April 1994. In these observations, we used the mosaic CCD camera
× ×
(Sekiguchi et al. 1992) which is made up of 15 1024 1024 CCD chips and covers 2 5
degree2 with 15 shots. The spatial resolution in the setup was 0.′′75 pix−1. Images were
taken in the R band with an exposure time of 20 minutes. The reduction of the data was
done by the usual method for optical imaging data. The weather condition during the
observation was neither stable nor photometric and the limiting magnitude changed from
′′
field to field. The typical seeing was 4 (FWHM) and the typical limiting magnitude was
∼
R = 21 mag, about one magnitude deeper than the APM data.
There are several optical objects within each error circle of X-ray source above R = 21
mag. Two X-ray sources (AX J132032+3326(227) and AX J130748+2925(002)) show clear
excesses of galaxies in and around their error circles (see Figure 1). These two objects have
been cataloged as candidate clusters of galaxies, Abell 1714 (Abell, Corwin, & Olowin 1989)
and Zwcl 1305.4+2941 (Zwicky, Herzog, & Wild 1961), respectively. Zwcl 1305.4+2941 was
also detected in the Einstein Medium Sensitivity Survey and its redshift was determined
to be z=0.241 (Stocke et al. 1991). We took deeper and better resolution images with a
– 9 –
× ′′
Tektronix 2048 2048 CCD on the University of Hawaii 88 telescope on 1995 March and
1996 April in the R and I band and confirmed the excess of galaxies. Thus, we identified
the two sources with clusters of galaxies. For other sources, active galaxies are the most
′
plausible counterparts. From optical objects within 0.8, which is slightly larger than the
′
estimated 90% confidence error radius (0.6), of each X-ray source, we selected targets for
spectroscopy, using 1) deeper X-ray follow-up data in the soft and hard band, 2) radio
emission, and 3) blueness of UV or optical color which indicate existence of an activity in a
object.
2.2.1. Deep Follow-up Observations in X-ray band
To pinpoint optical counterparts of X-ray sources, 20 ksec follow-up observations were
made with the ROSAT HRI in 2 fields in December 1997. We selected two fields centered
at α=13h14m36s, δ=32◦01′12′′ and α=13h12m29s, δ=31◦13′12′′ (J2000) with a radius of 19′,
to cover as many ASCA sources as possible. We summarize the data reduction and the
source detection in Appendix A. All of the X-ray sources in the SIS 2–7 keV 3.5σ sample
(AXJ131521+3159(136),AXJ131407+3158(127),andAXJ131327+3155(121)intheformer
field, and AX J131249+3112(096), AX J131128+3105(080), and AX J131321+3119(103) in
the latter field) except one (AX J131345+3118(104)) in these fields were detected by the
′′
HRI and their optical counterparts were pinpointed thanks to the 5 positional accuracy of
the HRI. On the position of the missed source (AX J131345+3118(104)), there is an X-ray
peak in the HRI image, though the significance level is slightly lower than the detection
limit. Since there is also an optical object at the position, the X-ray source is pinpointed.
It should be noted that there is no hard X-ray source which has an X-ray spectrum with an
apparent photon index smaller than 1.0 in these HRI observed fields.
∼
Deep 40 100 ksec pointing observations for 4 sources, which have hard X-ray
– 10 –
spectra, (AX J131501+3141(119), AX J131551+3237(171), AX J131210+3048(072), and
AX J130926+2952(016)) were made by ASCA and more precise positions and X-ray spectra
were obtained (Sakano et al. 1998; Sakano et al. 1999; Ueda et al. 1999b). The error circles
′
for these sources were estimated to be less than 0.6 at a 90% confidence level.
By the ROSAT PSPC, 15 and 13 ASCA LSS sources were detected in the ROSAT
PSPC All-Sky Survey (Voges et al. 1999) and pointing observations (Voges, private
communication), respectively. 9 of them were detected in both of the observations, thus
the X-ray positions of 19 ASCA sources were determined accurately. In details of the
cross-identification, see Appendix B.
2.2.2. Cross-correlation with FIRST radio source catalog
In the course of the optical counterpart selection, we examined the distribution of FIRST
radio sources around LSS X-ray sources to evaluate the cross-correlation between radio and
X-ray sources. The FIRST survey is a radio source survey conducted with the Very Large
Array in the 1.4 GHz band with a 5σ limiting flux of 1 mJy (Becker, White, & Helfand
′
1995). There are 17 radio sources within 0.5 from the centers of the X-ray sources; by
′ ′
contrast no radio source exists between 0.5 and 1 from them. Based on the surface number
density of detected radio sources in the LSS field, the contamination of a radio source which
is not an counterpart of an X-ray source is expected to be less than 1 source for the whole
′
34 X-ray sources within 0.5. Thus, the 17 radio sources are likely radio-counterparts of
X-ray sources. In details of the cross-identification, see Appendix B. Three X-ray sources
′
have two to four radio sources within 0.5. They are thought to be radio-loud objects with
radio lobes or clusters of galaxies. One of the three X-ray sources is already identified with
a cluster of galaxies (AX J132032+3326(227)). In summary, 35% (12/34) of the SIS 3.5σ
sample are detected in the FIRST survey and an optical counterpart is pinpointed thanks to